Risk evaluation through decision-support architectures in threat assessment and countering terrorism

• Published in: Applied soft computing 2011, Vol.11 (1), p.621-631
• Main Authors: Pedrycz, W., Chen, S.C., Rubin, S.H., Lee, G.
• Format: Article
• Language: English
• Subjects: Chem.-bio terrorism; Classification quality and risk assessment; Fuzzy clustering; Granulation of information; Shadowed sets; Threat detection; Three-valued logic
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2009.12.022

Owing to their inherent nature, terrorist activities could be highly diversified. The risk assessment becomes a crucial component as it helps us weigh pros and cons versus possible actions or some planning pursuits. The recognition of threats and their relevance/seriousness is an integral part of the overall process of classification, recognition, and assessing eventual actions undertaken in presence of acts of chem.-bio terrorism. In this study, we introduce an overall scheme of risk assessment realized on a basis of classification results produced for some experimental data capturing the history of previous threat cases. The structural relationships in these experimental data are first revealed with the help of information granulation – fuzzy clustering. We introduce two criteria using which information granules are evaluated, that is (a) representation capabilities which are concerned with the quality of representation of numeric data by abstract constructs such as information granules, and (b) interpretation aspects which are essential in the process of risk evaluation. In case of representation facet of information granules, we demonstrate how a reconstruction criterion quantifies their quality. Three ways in which interpretability is enhanced are studied. First, we show how to construct the information granules with extended cores (where the uncertainty associated with risk evaluation could be reduced) and shadowed sets, which provide a three-valued logic perspective of information granules given in the form of fuzzy sets. Subsequently, we show a way of interpreting fuzzy sets via an optimized set of its α-cuts.